Prediction of tropospheric ozone concentration using artificial neural networks at traffic and background urban locations in Novi Sad, Serbia.

In this paper, we described generation and performances of feedforward neural network models that could be used for a day ahead predictions of the daily maximum 1-h ozone concentration (1hO3) and 8-h average ozone concentration (8hO3) at one traffic and one background station in the urban area of Novi Sad, Serbia. The six meteorological variables for the day preceding the forecast and forecast day, ozone concentrations in the day preceding the forecast, the number of the day of the year, and the number of the weekday for which ozone prediction was performed were utilized as inputs. The three-layer perceptron neural network models with the best performance were chosen by testing with different numbers of neurons in the hidden layer and different activation functions. The mean bias error, mean absolute error, root mean squared error, correlation coefficient, and index of agreement or Willmott's Index for the validation data for 1hO3 forecasting were 0.005 µg m-3, 12.149 µg m-3, 15.926 µg m-3, 0.988, and 0.950, respectively, for the traffic station (Dnevnik), and - 0.565 µg m-3, 10.101 µg m-3, 12.962 µg m-3, 0.911, and 0.953, respectively, for the background station (Liman). For 8hO3 forecasting, statistical indicators were - 1.126 µg m-3, 10.614 µg m-3, 12.962 µg m-3, 0.910, and 0.948 respectively for the station Dnevnik and - 0.001 µg m-3, 8.574 µg m-3, 10.741 µg m-3, 0.936, and 0.966, respectively, for the station Liman. According to the Kolmogorov-Smirnov test, there is no significant difference between measured and predicted data. Models showed a good performance in forecasting days with the high values over a certain threshold.

Health risks of extended exposure to low-level UV radiation - An analysis of ground-based and satellite-derived data.

This study aims to indicate the importance of revising current health recommendations concerning the duration of exposure and individual sensitivity of the skin to solar ultraviolet (UV) radiation. For this purpose, a 16-year data series (2005-2020) of erythemal radiant exposure (Her) and UV index (UVI) for Serbia was analyzed. The UV-related risk was estimated for lighter skin (skin phototypes I-IV) under prolonged exposure on days when maximum UVI was below the recommended protection threshold (UVIlow days, for UVI < 3). Risk assessment was performed for seasonal exposure using satellite-derived data (OMUVBd product) previously validated by ground-based measurements in Novi Sad. The assessment of harmful effects included an analysis of the relation between the daily maximum UVI and the corresponding daily Her, the occurrence of UVIlow days, the exceedance of minimal erythema dose (MED), and the minimum duration of exposure to induce erythema (tMED) for all lighter skin phototypes. It was found that the share of UVIlow days in the total number of days in Serbia increases with the latitude, with the highest percentage in winter (up to 69.454%) and the lowest in summer (up to 3.468%). The results show that the daily Her frequently exceeded the harmful threshold for lighter skin phototypes I-IV (on average by 91.521, 84.923, 70.556, and 56.515%, respectively) on UVIlow days. It was found that prolonged exposure on days with a maximum of UVI = 2 poses a significant risk of erythema for all lighter skin phototypes, even for a duration of 3 h in the middle of the day, as well as medium risk for UVI = 1, and an absence of risk for UVI = 0. The results suggest that health recommendations should be revised, especially in the mid-latitudes, where the share of UVIlow days is large, and in areas where the population is predominantly lighter-skinned.